2 * Copyright (c) 1982, 1986, 1988, 1993
3 * The Regents of the University of California. All rights reserved.
5 * Redistribution and use in source and binary forms, with or without
6 * modification, are permitted provided that the following conditions
8 * 1. Redistributions of source code must retain the above copyright
9 * notice, this list of conditions and the following disclaimer.
10 * 2. Redistributions in binary form must reproduce the above copyright
11 * notice, this list of conditions and the following disclaimer in the
12 * documentation and/or other materials provided with the distribution.
13 * 3. All advertising materials mentioning features or use of this software
14 * must display the following acknowledgement:
15 * This product includes software developed by the University of
16 * California, Berkeley and its contributors.
17 * 4. Neither the name of the University nor the names of its contributors
18 * may be used to endorse or promote products derived from this software
19 * without specific prior written permission.
21 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
22 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
23 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
24 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
25 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
26 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
27 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
28 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
29 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
30 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
33 * @(#)ip_input.c 8.2 (Berkeley) 1/4/94
34 * ip_input.c,v 1.11 1994/11/16 10:17:08 jkh Exp
38 * Changes and additions relating to SLiRP are
39 * Copyright (c) 1995 Danny Gasparovski.
41 * Please read the file COPYRIGHT for the
42 * terms and conditions of the copyright.
54 static struct ip
*ip_reass(register struct ipasfrag
*ip
,
55 register struct ipq
*fp
);
56 static void ip_freef(struct ipq
*fp
);
57 static void ip_enq(register struct ipasfrag
*p
,
58 register struct ipasfrag
*prev
);
59 static void ip_deq(register struct ipasfrag
*p
);
62 * IP initialization: fill in IP protocol switch table.
63 * All protocols not implemented in kernel go to raw IP protocol handler.
68 ipq
.next
= ipq
.prev
= (ipqp_32
)&ipq
;
69 ip_id
= tt
.tv_sec
& 0xffff;
75 * Ip input routine. Checksum and byte swap header. If fragmented
76 * try to reassemble. Process options. Pass to next level.
82 register struct ip
*ip
;
85 DEBUG_CALL("ip_input");
86 DEBUG_ARG("m = %lx", (long)m
);
87 DEBUG_ARG("m_len = %d", m
->m_len
);
89 STAT(ipstat
.ips_total
++);
91 if (m
->m_len
< sizeof (struct ip
)) {
92 STAT(ipstat
.ips_toosmall
++);
96 ip
= mtod(m
, struct ip
*);
98 if (ip
->ip_v
!= IPVERSION
) {
99 STAT(ipstat
.ips_badvers
++);
103 hlen
= ip
->ip_hl
<< 2;
104 if (hlen
<sizeof(struct ip
) || hlen
>m
->m_len
) {/* min header length */
105 STAT(ipstat
.ips_badhlen
++); /* or packet too short */
109 /* keep ip header intact for ICMP reply
110 * ip->ip_sum = cksum(m, hlen);
114 STAT(ipstat
.ips_badsum
++);
119 * Convert fields to host representation.
122 if (ip
->ip_len
< hlen
) {
123 STAT(ipstat
.ips_badlen
++);
130 * Check that the amount of data in the buffers
131 * is as at least much as the IP header would have us expect.
132 * Trim mbufs if longer than we expect.
133 * Drop packet if shorter than we expect.
135 if (m
->m_len
< ip
->ip_len
) {
136 STAT(ipstat
.ips_tooshort
++);
139 /* Should drop packet if mbuf too long? hmmm... */
140 if (m
->m_len
> ip
->ip_len
)
141 m_adj(m
, ip
->ip_len
- m
->m_len
);
143 /* check ip_ttl for a correct ICMP reply */
144 if(ip
->ip_ttl
==0 || ip
->ip_ttl
==1) {
145 icmp_error(m
, ICMP_TIMXCEED
,ICMP_TIMXCEED_INTRANS
, 0,"ttl");
150 * Process options and, if not destined for us,
151 * ship it on. ip_dooptions returns 1 when an
152 * error was detected (causing an icmp message
153 * to be sent and the original packet to be freed).
155 /* We do no IP options */
156 /* if (hlen > sizeof (struct ip) && ip_dooptions(m))
160 * If offset or IP_MF are set, must reassemble.
161 * Otherwise, nothing need be done.
162 * (We could look in the reassembly queue to see
163 * if the packet was previously fragmented,
164 * but it's not worth the time; just let them time out.)
166 * XXX This should fail, don't fragment yet
168 if (ip
->ip_off
&~ IP_DF
) {
169 register struct ipq
*fp
;
171 * Look for queue of fragments
174 for (fp
= (struct ipq
*) ipq
.next
; fp
!= &ipq
;
175 fp
= (struct ipq
*) fp
->next
)
176 if (ip
->ip_id
== fp
->ipq_id
&&
177 ip
->ip_src
.s_addr
== fp
->ipq_src
.s_addr
&&
178 ip
->ip_dst
.s_addr
== fp
->ipq_dst
.s_addr
&&
179 ip
->ip_p
== fp
->ipq_p
)
185 * Adjust ip_len to not reflect header,
186 * set ip_mff if more fragments are expected,
187 * convert offset of this to bytes.
190 if (ip
->ip_off
& IP_MF
)
191 ((struct ipasfrag
*)ip
)->ipf_mff
|= 1;
193 ((struct ipasfrag
*)ip
)->ipf_mff
&= ~1;
198 * If datagram marked as having more fragments
199 * or if this is not the first fragment,
200 * attempt reassembly; if it succeeds, proceed.
202 if (((struct ipasfrag
*)ip
)->ipf_mff
& 1 || ip
->ip_off
) {
203 STAT(ipstat
.ips_fragments
++);
204 ip
= ip_reass((struct ipasfrag
*)ip
, fp
);
207 STAT(ipstat
.ips_reassembled
++);
217 * Switch out to protocol's input routine.
219 STAT(ipstat
.ips_delivered
++);
222 tcp_input(m
, hlen
, (struct socket
*)NULL
);
231 STAT(ipstat
.ips_noproto
++);
241 * Take incoming datagram fragment and try to
242 * reassemble it into whole datagram. If a chain for
243 * reassembly of this datagram already exists, then it
244 * is given as fp; otherwise have to make a chain.
247 ip_reass(register struct ipasfrag
*ip
, register struct ipq
*fp
)
249 register struct mbuf
*m
= dtom(ip
);
250 register struct ipasfrag
*q
;
251 int hlen
= ip
->ip_hl
<< 2;
254 DEBUG_CALL("ip_reass");
255 DEBUG_ARG("ip = %lx", (long)ip
);
256 DEBUG_ARG("fp = %lx", (long)fp
);
257 DEBUG_ARG("m = %lx", (long)m
);
260 * Presence of header sizes in mbufs
261 * would confuse code below.
262 * Fragment m_data is concatenated.
268 * If first fragment to arrive, create a reassembly queue.
272 if ((t
= m_get()) == NULL
) goto dropfrag
;
273 fp
= mtod(t
, struct ipq
*);
275 fp
->ipq_ttl
= IPFRAGTTL
;
276 fp
->ipq_p
= ip
->ip_p
;
277 fp
->ipq_id
= ip
->ip_id
;
278 fp
->ipq_next
= fp
->ipq_prev
= (ipasfragp_32
)fp
;
279 fp
->ipq_src
= ((struct ip
*)ip
)->ip_src
;
280 fp
->ipq_dst
= ((struct ip
*)ip
)->ip_dst
;
281 q
= (struct ipasfrag
*)fp
;
286 * Find a segment which begins after this one does.
288 for (q
= (struct ipasfrag
*)fp
->ipq_next
; q
!= (struct ipasfrag
*)fp
;
289 q
= (struct ipasfrag
*)q
->ipf_next
)
290 if (q
->ip_off
> ip
->ip_off
)
294 * If there is a preceding segment, it may provide some of
295 * our data already. If so, drop the data from the incoming
296 * segment. If it provides all of our data, drop us.
298 if (q
->ipf_prev
!= (ipasfragp_32
)fp
) {
299 i
= ((struct ipasfrag
*)(q
->ipf_prev
))->ip_off
+
300 ((struct ipasfrag
*)(q
->ipf_prev
))->ip_len
- ip
->ip_off
;
311 * While we overlap succeeding segments trim them or,
312 * if they are completely covered, dequeue them.
314 while (q
!= (struct ipasfrag
*)fp
&& ip
->ip_off
+ ip
->ip_len
> q
->ip_off
) {
315 i
= (ip
->ip_off
+ ip
->ip_len
) - q
->ip_off
;
322 q
= (struct ipasfrag
*) q
->ipf_next
;
323 m_freem(dtom((struct ipasfrag
*) q
->ipf_prev
));
324 ip_deq((struct ipasfrag
*) q
->ipf_prev
);
329 * Stick new segment in its place;
330 * check for complete reassembly.
332 ip_enq(ip
, (struct ipasfrag
*) q
->ipf_prev
);
334 for (q
= (struct ipasfrag
*) fp
->ipq_next
; q
!= (struct ipasfrag
*)fp
;
335 q
= (struct ipasfrag
*) q
->ipf_next
) {
336 if (q
->ip_off
!= next
)
340 if (((struct ipasfrag
*)(q
->ipf_prev
))->ipf_mff
& 1)
344 * Reassembly is complete; concatenate fragments.
346 q
= (struct ipasfrag
*) fp
->ipq_next
;
349 q
= (struct ipasfrag
*) q
->ipf_next
;
350 while (q
!= (struct ipasfrag
*)fp
) {
353 q
= (struct ipasfrag
*) q
->ipf_next
;
358 * Create header for new ip packet by
359 * modifying header of first packet;
360 * dequeue and discard fragment reassembly header.
361 * Make header visible.
363 ip
= (struct ipasfrag
*) fp
->ipq_next
;
366 * If the fragments concatenated to an mbuf that's
367 * bigger than the total size of the fragment, then and
368 * m_ext buffer was alloced. But fp->ipq_next points to
369 * the old buffer (in the mbuf), so we must point ip
370 * into the new buffer.
372 if (m
->m_flags
& M_EXT
) {
374 delta
= (char *)ip
- m
->m_dat
;
375 ip
= (struct ipasfrag
*)(m
->m_ext
+ delta
);
378 /* DEBUG_ARG("ip = %lx", (long)ip);
379 * ip=(struct ipasfrag *)m->m_data; */
383 ((struct ip
*)ip
)->ip_src
= fp
->ipq_src
;
384 ((struct ip
*)ip
)->ip_dst
= fp
->ipq_dst
;
386 (void) m_free(dtom(fp
));
388 m
->m_len
+= (ip
->ip_hl
<< 2);
389 m
->m_data
-= (ip
->ip_hl
<< 2);
391 return ((struct ip
*)ip
);
394 STAT(ipstat
.ips_fragdropped
++);
400 * Free a fragment reassembly header and all
401 * associated datagrams.
404 ip_freef(struct ipq
*fp
)
406 register struct ipasfrag
*q
, *p
;
408 for (q
= (struct ipasfrag
*) fp
->ipq_next
; q
!= (struct ipasfrag
*)fp
;
410 p
= (struct ipasfrag
*) q
->ipf_next
;
415 (void) m_free(dtom(fp
));
419 * Put an ip fragment on a reassembly chain.
420 * Like insque, but pointers in middle of structure.
423 ip_enq(register struct ipasfrag
*p
, register struct ipasfrag
*prev
)
425 DEBUG_CALL("ip_enq");
426 DEBUG_ARG("prev = %lx", (long)prev
);
427 p
->ipf_prev
= (ipasfragp_32
) prev
;
428 p
->ipf_next
= prev
->ipf_next
;
429 ((struct ipasfrag
*)(prev
->ipf_next
))->ipf_prev
= (ipasfragp_32
) p
;
430 prev
->ipf_next
= (ipasfragp_32
) p
;
434 * To ip_enq as remque is to insque.
437 ip_deq(register struct ipasfrag
*p
)
439 ((struct ipasfrag
*)(p
->ipf_prev
))->ipf_next
= p
->ipf_next
;
440 ((struct ipasfrag
*)(p
->ipf_next
))->ipf_prev
= p
->ipf_prev
;
444 * IP timer processing;
445 * if a timer expires on a reassembly
451 register struct ipq
*fp
;
453 DEBUG_CALL("ip_slowtimo");
455 fp
= (struct ipq
*) ipq
.next
;
461 fp
= (struct ipq
*) fp
->next
;
462 if (((struct ipq
*)(fp
->prev
))->ipq_ttl
== 0) {
463 STAT(ipstat
.ips_fragtimeout
++);
464 ip_freef((struct ipq
*) fp
->prev
);
470 * Do option processing on a datagram,
471 * possibly discarding it if bad options are encountered,
472 * or forwarding it if source-routed.
473 * Returns 1 if packet has been forwarded/freed,
474 * 0 if the packet should be processed further.
483 register struct ip
*ip
= mtod(m
, struct ip
*);
485 register struct ip_timestamp
*ipt
;
486 register struct in_ifaddr
*ia
;
487 /* int opt, optlen, cnt, off, code, type = ICMP_PARAMPROB, forward = 0; */
488 int opt
, optlen
, cnt
, off
, code
, type
, forward
= 0;
489 struct in_addr
*sin
, dst
;
490 typedef u_int32_t n_time
;
494 cp
= (u_char
*)(ip
+ 1);
495 cnt
= (ip
->ip_hl
<< 2) - sizeof (struct ip
);
496 for (; cnt
> 0; cnt
-= optlen
, cp
+= optlen
) {
497 opt
= cp
[IPOPT_OPTVAL
];
498 if (opt
== IPOPT_EOL
)
500 if (opt
== IPOPT_NOP
)
503 optlen
= cp
[IPOPT_OLEN
];
504 if (optlen
<= 0 || optlen
> cnt
) {
505 code
= &cp
[IPOPT_OLEN
] - (u_char
*)ip
;
515 * Source routing with record.
516 * Find interface with current destination address.
517 * If none on this machine then drop if strictly routed,
518 * or do nothing if loosely routed.
519 * Record interface address and bring up next address
520 * component. If strictly routed make sure next
521 * address is on directly accessible net.
525 if ((off
= cp
[IPOPT_OFFSET
]) < IPOPT_MINOFF
) {
526 code
= &cp
[IPOPT_OFFSET
] - (u_char
*)ip
;
529 ipaddr
.sin_addr
= ip
->ip_dst
;
530 ia
= (struct in_ifaddr
*)
531 ifa_ifwithaddr((struct sockaddr
*)&ipaddr
);
533 if (opt
== IPOPT_SSRR
) {
535 code
= ICMP_UNREACH_SRCFAIL
;
539 * Loose routing, and not at next destination
540 * yet; nothing to do except forward.
544 off
--; / * 0 origin
* /
545 if (off
> optlen
- sizeof(struct in_addr
)) {
547 * End of source route. Should be for us.
549 save_rte(cp
, ip
->ip_src
);
553 * locate outgoing interface
555 bcopy((caddr_t
)(cp
+ off
), (caddr_t
)&ipaddr
.sin_addr
,
556 sizeof(ipaddr
.sin_addr
));
557 if (opt
== IPOPT_SSRR
) {
558 #define INA struct in_ifaddr *
559 #define SA struct sockaddr *
560 if ((ia
= (INA
)ifa_ifwithdstaddr((SA
)&ipaddr
)) == 0)
561 ia
= (INA
)ifa_ifwithnet((SA
)&ipaddr
);
563 ia
= ip_rtaddr(ipaddr
.sin_addr
);
566 code
= ICMP_UNREACH_SRCFAIL
;
569 ip
->ip_dst
= ipaddr
.sin_addr
;
570 bcopy((caddr_t
)&(IA_SIN(ia
)->sin_addr
),
571 (caddr_t
)(cp
+ off
), sizeof(struct in_addr
));
572 cp
[IPOPT_OFFSET
] += sizeof(struct in_addr
);
574 * Let ip_intr's mcast routing check handle mcast pkts
576 forward
= !IN_MULTICAST(ntohl(ip
->ip_dst
.s_addr
));
580 if ((off
= cp
[IPOPT_OFFSET
]) < IPOPT_MINOFF
) {
581 code
= &cp
[IPOPT_OFFSET
] - (u_char
*)ip
;
585 * If no space remains, ignore.
588 if (off
> optlen
- sizeof(struct in_addr
))
590 bcopy((caddr_t
)(&ip
->ip_dst
), (caddr_t
)&ipaddr
.sin_addr
,
591 sizeof(ipaddr
.sin_addr
));
593 * locate outgoing interface; if we're the destination,
594 * use the incoming interface (should be same).
596 if ((ia
= (INA
)ifa_ifwithaddr((SA
)&ipaddr
)) == 0 &&
597 (ia
= ip_rtaddr(ipaddr
.sin_addr
)) == 0) {
599 code
= ICMP_UNREACH_HOST
;
602 bcopy((caddr_t
)&(IA_SIN(ia
)->sin_addr
),
603 (caddr_t
)(cp
+ off
), sizeof(struct in_addr
));
604 cp
[IPOPT_OFFSET
] += sizeof(struct in_addr
);
608 code
= cp
- (u_char
*)ip
;
609 ipt
= (struct ip_timestamp
*)cp
;
610 if (ipt
->ipt_len
< 5)
612 if (ipt
->ipt_ptr
> ipt
->ipt_len
- sizeof (int32_t)) {
613 if (++ipt
->ipt_oflw
== 0)
617 sin
= (struct in_addr
*)(cp
+ ipt
->ipt_ptr
- 1);
618 switch (ipt
->ipt_flg
) {
620 case IPOPT_TS_TSONLY
:
623 case IPOPT_TS_TSANDADDR
:
624 if (ipt
->ipt_ptr
+ sizeof(n_time
) +
625 sizeof(struct in_addr
) > ipt
->ipt_len
)
627 ipaddr
.sin_addr
= dst
;
628 ia
= (INA
)ifaof_ i f p
foraddr((SA
)&ipaddr
,
632 bcopy((caddr_t
)&IA_SIN(ia
)->sin_addr
,
633 (caddr_t
)sin
, sizeof(struct in_addr
));
634 ipt
->ipt_ptr
+= sizeof(struct in_addr
);
637 case IPOPT_TS_PRESPEC
:
638 if (ipt
->ipt_ptr
+ sizeof(n_time
) +
639 sizeof(struct in_addr
) > ipt
->ipt_len
)
641 bcopy((caddr_t
)sin
, (caddr_t
)&ipaddr
.sin_addr
,
642 sizeof(struct in_addr
));
643 if (ifa_ifwithaddr((SA
)&ipaddr
) == 0)
645 ipt
->ipt_ptr
+= sizeof(struct in_addr
);
652 bcopy((caddr_t
)&ntime
, (caddr_t
)cp
+ ipt
->ipt_ptr
- 1,
654 ipt
->ipt_ptr
+= sizeof(n_time
);
665 /* ip->ip_len -= ip->ip_hl << 2; XXX icmp_error adds in hdr length */
668 icmp_error(m
, type
, code
, 0, 0);
670 STAT(ipstat
.ips_badoptions
++);
677 * Strip out IP options, at higher
678 * level protocol in the kernel.
679 * Second argument is buffer to which options
680 * will be moved, and return value is their length.
681 * (XXX) should be deleted; last arg currently ignored.
684 ip_stripoptions(m
, mopt
)
685 register struct mbuf
*m
;
689 struct ip
*ip
= mtod(m
, struct ip
*);
690 register caddr_t opts
;
693 olen
= (ip
->ip_hl
<<2) - sizeof (struct ip
);
694 opts
= (caddr_t
)(ip
+ 1);
695 i
= m
->m_len
- (sizeof (struct ip
) + olen
);
696 memcpy(opts
, opts
+ olen
, (unsigned)i
);
699 ip
->ip_hl
= sizeof(struct ip
) >> 2;